Feb. 20, 1964
COMMUNICATIONS TO THE EDITOR
integration with the aid of an internal standard reveals the presence of 8.0 f 0.5 fluorines. The position of this peak is in complete accord with that expected for the fluoroborate anion : thus, triphenylmethyl fluoroborate in methylene chloride appears a t 71.7 p.p.m. and aqueous fluoroboric acid is a t 71.3 p.p.m.' For this data to be consistent with the presence of V as the major species, a rapid time-averaged exchange of the ring fluorine, boron trifluoride, and fluoroborate anion is required. Since the boron trifluoride resonance is found a t +47.8 p.p.m. in methylene chloride, the fortuitous appearance of the exchange-averaged resonance a t the exact chemical shift of authentic fluoroborate dictates that the ring fluorine of V, in the absence of exchange, appear a t +140 p.p.m., a region in which apparently only aliphatic primary fluorine has been observed.8 The absence of Flg-Bl1 splittingg in the spectrum of IV could be interpreted as evidence for the presence of exchange averaging. However, this splitting is also absent in the trityl fluoroborate spectrum in which rapid equilibration of chemically different fluorines (as required for V) does not occur. At low temperatures the peak in question does broaden, perhaps due to a restoration of the F19-B11splitting. It now appears that differentiation between monoand dication, as for example I and 11, by its proton n.m.r. spectrum cannot be accomplished on an a priori basis, and details of these investigations will be reported in due course. Acknowledgment.-We are particularly indebted to Dr. V. R. Sandel for many helpful discussions.
735
+
(7) G . A. Olah, e t al., J . A m . Chem. SOL.,84,2733 (1962). (8) W. D. Phillips, "Determinations of Organic Structures by Physical Methods," Academic Press, New York, N . Y . , 1962,p. 406. (9) F19-B11 splitting of the B F I - resonance into a poorly resolved quartet has been observed in a concentrated aqueous solution of h7aBFd by R. D. Chambers, el al., Can J . Chem., 39,258 (1961).
1
Fig. 1.-The
A Fig. 2.-The
I
,
,
9
$0".
n.m.r. spectrum of 11.
A
n.m.r. spectrum of the lac,satellite of I1 ( A ) ; added ethanol ( B ) ; spinning side bands ( C ) .
sided within 2 hr. and silicic acid chromatography of the photolysate easily separated a new ketone (11), m.p. 139-140' dec. (Anal. Calcd. for C17H120:C, 87.90; H , 5.21; mol. wt., 232. Found: C, 87.88; H, 5.39 ; mol. wt., 230 (thermoelectric osmometer)). H
EASTERN RESEARCH LABORATORY H. H. FREEDMAN THEDow CHEMICAL COMPAKY A. E. YOUNG FRAMINGHAM, MASSACHUSETTS RECEIVED OCTOBER21, 1963
Strained Systems. I.
The Tricyclo [1.1. l.0435]pentane System
31
CHN~ H I I1 Sir : While the infrared spectrum indicated the presence of a It is well known that the photolysis of an a-diazofour-membered ketone (1785 cm. - l ) , the n.m.r. specmethyl ketone effects a reaction analogous to the Wolff trum in CDC13 (Fig. 1) showed two equivalent aliphatic rearrangement.' We have found, however, that an protons a t 7.12 r and a phenyl absorption a t 2.65 T a-ketocarbene generated by ultraviolet irradiation (10 protons). These analytical and spectral results undergoes an intramolecular addition to a double are in complete agreement with the structure shown bond a t a greater rate than the rearrangement.* Thus, the photolysis of 1,2-diphenyl-3-diazomethylketocyclo- in 11. Catalytic hydrogenation of I1 proceeded with the propene (I) has led to the formation of a hitherto unuptake of tw8 moles of hydrogen and provided in excelknown ring system. We now wish to report the lent yield cis-3,4-diphenylcyclopentanone(111),5 m.p. synthesis and properties of this extremely strained 106-107", identified by direct comparison with an tricyclo [ 1.1.1.04,S]pentane system. authentic sample. Since the bicyclobutane system is Compound I , m.p. 115' dec., was prepared from the 0 corresponding acid chloride, m.p. 101-102'. A 0.5% solution of I in tetrahydrofuran was irradiated a t room temperature with a Hanovia 450-w. mercury lamp, using a Pyrex f i l t e ~ - . ~Evolution ,~ of nitrogen sub(1) P. de Mayo, "Advances in Organic Chemistry, Methods and Results,'' Vol 11, Interscience Publishers, Inc., New York, N . Y., 1960,p. 407. For compariqon with t h e silver-catalyzed rearrangement, see K. B . Wiberg and T . W . H u t t o n , J . A m . Chem Soc., 78, 1640 (1956). ( 2 ) Addition of a n u-ketocarbene t o a double bond is also effected with copper-catalyzed thermal decomposition of the diazomethyl ketone. For instance, see G . Stork and J. Ficin;, ibid., 8 3 , 4678 (1961). However, in cases where the products are thermally unstable, this method is not applicable. (3) Treatment of I with silver benzoate in triethylamine effected t h e Wolff rearrangement and provided t h e methyl ester of t h e homologated acid in good yield (4) T h e photolysis of 1,2-di-n-propyl-3-diazomethylketocyclopropene provided a ketone which corresponds t o 11. T h e n.m.r. spectrum of this
I11 known to undergo hydrogeno1ysisj6 the formation of di-n-propyl derivative showed a singlet (2 protons) a t 8.13 T in addition t o a typical n-propyl absorption. This experiment was carried out by M r . N . T . Castellucci. (5) D.Y . Curtin, et a i . , J . A m . Chem. Soc., 83, 4838 (1961),and references cited therein. (6) K . R. Wiberg and R . P. Ciula, ibid.,81,5261 (1959); J. Meinwald, C . Swithenhank, and A. Lewis, ibid., 8 6 , 1880 (1963). T h e course of hydrogenolysis of I1 is somewhat different from those of l-carboethoxybicyclo-
736
Vol. 86
COMMUNICATIONS TO THE EDITOR
I11 from I1 presents strong evidence in support of assigned structure 11. The n.m.r. spectrum7 (Fig. 2) of the satellite of compound I1 revealed remarkably high coupling constants both for ' T - H and H1-H3 (I3C1-H = 190.1 c.P.s., H1-Ha = 14 c.P.s.). Using an equation to correlate coupling constants with the extent of sp hybridization of the carbon atomic orbitals used in the bonds,8 we obtain 38% s-character for the exocyclic orbital of the C1 atom. This high degree of s-character is a clear demonstration of the great strain involved in this ring system. In view of the fact that long-range proton spin-spin coupling constants are in the order of 0.3-0.5 C.P.S. with methyl ketones,I0the size of the coupling constant observed for H1-Ha of I1 is remarkable. Although the theoretical aspect of the long-range spin-spin interaction still remains unclear, this large coupling evidently arises from the unique geometrical arrangement of the two hydrogens and two carbon atoms, i . e . , these four atoms lie in a straight line and the distance of the carbon atoms is considerably shorter than 2.7 fL. Under these circumstances, an appreciable interaction of the two C-H orbitals (in particular, because of their large s-character) seems reasonable. The ultraviolet spectrum of I1 in methanol showed a maximum a t 242 mp (E 1.46 X lo4). Thus a distinctive conjugation effect of cyclopropanes is observable in the case of the bicyclobutane system. We are observing unique reactions intrinsic to the bicyclobutane system as well as unusual solvolytic behaviors of the tosylate of 4,5-diphenyltricyclo[1.1.1.04*5]pentan-2-ol. These results will be reported in a separate paper. Acknowledgment.-The author is grateful to Dr. A . A . Bothner-By for helpful discussions and Mr. N. T. Castellucci for his skillful assistance.
ene-3,17-dione (VI)5,6was incubated with the same microorganism (see Scheme I, arabic numbers indicate the chemical shifts ( 7 ) in deuteriochloroform solutions), Therefore, one of the two hydroxyl groups in each compound should be situated a t the Sa-position. Compound 11 [m.p. 224-226", [ a I z 4 $144.2' ~ (dioxane), Xmax 242 mp (E 15,400), 3441, 1724, 1659, and 1617 ~ m . - ~ ]and ' its monoacetate (VII, sirup, v",",': 3466, 1745, 1674, 1622, and 1240 cm.-l) show no signal corresponding to the 18-methyl group in their n.m.r. spectra.8 Instead, oxygen-bearing methylene signals (about 6.28 7 in I1 and about 5.73 7 in VII) are observed. The 19-methyl signal of I1 appears a t 8.66 7 shifted by -0.12 p.p.m. from that of I , owing to the Sa-hydroxyl g r o ~ p . ~ ,Chromium '~ trioxide oxidation of I1 afforded an aldehyde [VIII, m.p. 208-211°, [ a I z 5 D -7.3", Xmax 240.5 mp (E 15,800), vgl:"' 3400, 1747, 1692, 1646, and 1616 cm.-l], which was verified by the appearance of the n.m.r. signal of an aldehydic proton a t 0.47 7 . These results suggest that I1 is the 18-hydroxyl derivative of VI. Confirmatory evidence that I1 has an 18-hydroxyl group, not a 19-hydroxyl group, was obtained by the following degradation of I1 with Bacillus sphuericus (ATCC 7055) having the ability to 1-dehydrogenate various steroids. I 1 - l 3 Microbiological dehydrogenation of I1 afforded an expected 9, IO-seco compound [IX, m.p. 148-150°, [ a ] 2 7 . 5 D $ 9 l 0 , Xmax 279.5 mp (E 2350), vzip' 3452, 1727, 1693, 1611, and 1512 cm.-l]. The n.m.r. spectra of I X and its diacetate (X, sirup, v:,",': 1751, 1713, 1612, 1591, 1502, and 1213 cm.-I) show the signal of a methyl group attached to the benzene ring a t 7.76 and 7.68 7, respectively, and an ABtype quartet due to the oxygen-bearing methylene group. The angular methyl signal that appears in TI is not observed in I X and X. For the purpose of comparison, 9,10-seco-3-hydroxyandrosta1,3,5( 10)-triene-9,17-di0ne~,'~ (XI, m.p. 122-123', [ a I z 7 D $96') [ l .l.O]butane and tricyclo [Z.1.1.0~~~lhexane-5-l-butylcarboxamide. However, this difference is easily rationalized by t h e steric hindrance of t h e two and its acetate (XII, m.p. 145-146", [ L Y ] ~ ' . ~ D$83') phenyl groups of 11. were prepared by the same method. The main n.m.r. (7) T h e author i s grateful t o Mr. Joey Holcombe for determining this features of X I and XI1 resemble those of I X and X , n.m.r. spectrum with a Varian Associates HR-100 spectrometer. respectively, except that the 18-methyl signals in X I (8) N. Muller and I). E. Prltchard, J. Chem. Phys.. 81, 1471 (1959). and XI1 are replaced by a doublet of doublets in I X (9) G. L. Closs and L. E . Closs, J A m . Chem. SOL.,85, 2022 (1963). (IO) N. van Meurs, Speclvochim A c t o , 19, 1695 (1963), and references and X . Thus the structure of I1 was established to cited therein, B. L Shapiro. unpublished be 9 a ,18-dihydroxyandrost-4-ene-3,17-dione. MELLONISSTITUTE SATORU MASAMUNE Compound 111 [m.p. 253-257", [ a ] 2 3 . 5 D $213.8', PITTSBURGH, PENXSYLVANIA Xmax 241 mp, (E 16,200), v 2 : O ' 3434, 3359, 1720, 1678, RECEIVED JASUARY 2, 1964 and 1625 cm.-l] was converted into its monoacetate $201.3', h m a x 241 mM [XIII, m.p. 254-256', Microbiological 9a,12 a- and 9a,18-Dihydroxylations ( E 16,000),v::?' 3565, 1740, 1661, 1617, and 1240 cm.-'] by the usual acetylation. Chromium trioxide oxidaof Androst-Cene-3,17-dione
Sir : Although a large number of papers dealing with the microbiological hydroxylation of steroids have appeared in recent years, the 18-hydroxylation has not been reported so far. We wish to report here the successful hydroxylation a t the 9a-, 12a-, 14a-, l+, and 18-positions of androst-4-ene-3,17-dione (I). Among these, the 18-hydroxylation is believed to be the most important reaction because of its usefulness in preparing such 18-oxygenated steroids as aldosterone. Transformation of I with the use of Cercospora melonis [ Corynespora melonis (Cke) Lindau ] produced four dihydroxylated compounds (11, 111, IV, and V) by the method previously d e ~ c r i b e d . ~ These four products were obtained also when 9a-hydroxyandrost-4(1) B y t h e use of this microorganism, Reichstein's substance S is converted into its 88- and l,5@-hydroxyl derivatives (see ref. 2 and 3, respecti vely I . ( 2 ) K . Tori and E. Kondo, Telruhedvon L e l l e r s , No. 10, 64.5 (1963). (3) E. Kondo. t o be published. (4) E. Kondo, K. Morihara, Y. Nozaki, and E. Masuo, J. A g r . Chem. SOL. J a p a i i , 34. 844 (1960).
vr:rl
(5) E. Kondo and T. Mitsugi, ibid., 36, 521 (1961). (6) R . M . Dodson and R. D. Muir, J . A m . Chem. SOL.,80, 6148 (1958). (7) Elementary analyses of t h e compounds described here gave satisfactory values Unless otherwise noted, optical rotations were determined in chloroform containing 1% ethanol, and ultraviolet absorption spectra were observed on t h e solutions in 95% ethanol. (8) All n.m.r spectra were taken with a Varian A-60 spectrometer on 2-5% solutions in deuteriochloroform containing tettamethylsilane a s an internal reference a t room temperature, Chemical shifts are expressed in r-values. (9) T h e additivity rule for signal shifts of t h e angular methyl groups due t o shielding by various functional groups has been well established [for example, see R . F. Ziircher, Helv. Chrm. A c l a , 44, 1380 (1961)l. (10) T h e shift value of the 19-methyl due t o a Qa-hydroxyl group was reported t o he about -0.14 p.p.m.2 I t should be noted t h a t t h e signal shift of the 4-proton by about -0.15 p . p . m . is characteristic for the introduction of a Sa-hydroxyl group into a A4-3-ketosteroid. (11) T. H . Stoudt, W. J . McAleer, J. M. Chemerda, M . A. Kozlowski, R . F. Hirschmann, V. Marlatt, and R. Miller, Arch. Biochem. B i o p h y s . , 59, 304 (1955). (12) E . Kondo and E . Masuo, A n n . Repl. Shionogi Res. Lab., 10, 103 (1960) (13) I t has been well known t h a t 9a-hydroxy-A4-3-keto steroids are transformed t o their 9.10-secophenol derivatives by microorganisms capable of introducing a AI-double bond.8." (14) C. J. Sih Biochim. Biophys Acta, 62, 541 (1962).